This alog follows up LHO:81769 where I calibrated the ASC drives to test mass motion for all eight arm ASC control loops. Now, I have taken the noise budget injections that we run to measure the ASC coupling to DARM, and used that to calibrate an angle-to-length coupling function in mm/rad. I have only done this for the HARD loops because the SOFT loops do not couple very strongly to DARM (notable exception to CSOFT P, which I will follow up on).

The noise budget code uses an excess power projection to DARM, but instead I chose to measure the linear transfer function. The coherence is just ok, so I think a good follow up is to remeasure the coupling again and drive a bit harder/average longer (these are 60 second measurements). This plot shows the noise budget injection into calibrated DARM/ASC PUM drive [m/Nm] transfer function, and the coherence of the measurement.

I followed a similar calibration procedure to my previous alog:

• Go from "H1:ASC-[DOF]_[P/Y]_OUT_DQ"  to Nm of PUM torque using G1100968
  • I made sure that I corrected by an erroneous factor of two that I found previously
  • Some of these injections were done after the ETMX DAC change. However, Daniel noted this alog for me, 80023, so the calibration for the ASC should not change significantly.
•  transform this N*m PUM drive to radians of test mass motion using the PUM-TST state space suspension model transfer function, which is calibrated into real units (radian/Nm)
  • I did all of this in python, so I used Kevin's exported quad state space model, found here: T2300299
  • I used the damped model which applies the ETMX top mass damping filters
• Multiply by 1e3 to convert to mm/rad

I did not apply the drive matrix here, so the calibration is into ETM motion only (factor of +/-1), whereas the calibration into ITM motion would have an additional +/- 0.74 (+/- 0.72 for yaw) applied.

HARD Pitch Angle to Length coupling plot

HARD Yaw Angle to Length coupling plot

Overall, the best-measured DOF here is CHARD Y. In both CHARD Y and DHARD Y, there seem to be two clear coupling regions: one fairly flat region above 20 Hz in CHARD Y and above 30 Hz in DHARD Y, reaching between 20-30 mm/rad. Below, there is a steep coupling. This is reminiscient of the coupling that Gabriele, Louis, and I measured back in March and tried to mitigate with A2L and WFS offset. We found that we could reduce the flatter coupling in DHARD Y by adjusting the A2L gain, and the steeper coupling by applying a small offset in AS WFS A yaw DC. We are currently not running with that WFS offset. The yaw coupling suggests that we have some sort of miscentering on both the REFL and AS WFS which causes a steep low frequency coupling which is less sensitive to beam centering on the test mass (as shown by the A2L tests); meanwhile, the flat coupling is sensitive to beam miscentering on the test mass, which is expected (see e.g. T0900511).

The pitch coupling has the worst coherence here, but the coupling is certainly not flat. It appears to be rising with about f^4 at high frequency. I have a hard time understanding what could cause that. There is also possibly a similar steep coupling at low frequency like the yaw coupling, but the coherence is so poor it's hard to see.

Assuming that I have my calibration factors correct here (please don't assume this! check my work!), this suggests that the beam miscentering is higher than 1 mm everywhere and possibly up to 30 mm on the ETMs (remember this would be ~25% lower on the ITMs). This seems very large, so I'm hoping that there is another errant factor of two or something somewhere.

My code for both the calibrated motion and calibrated coupling is in a git repo here: https://git.ligo.org/ecapote/ASC_calibration